Yusuke Oda
Project Associate Professor
Chiyoda, Japan
Summary
🤩
Rockstar🎓
Top SchoolYusuke Oda is a project associate professor and technical director with 11 years of industry and research experience building machine learning systems and developer tooling. He has worked at Google on Translate and Brain, contributed to TensorFlow's Lingvo framework (notably refining GMM Attention), and is the primary author of the popular latexify_py library that translates Python code into LaTeX. Based in Chiyoda, Japan, he bridges academia and industry through roles at NII, NAIST, RIKEN and Cierpa, applying research-grade methods to production software and pipelines. His background spans deep learning research, translation systems, and back-end engineering, with practical expertise in reshaping data pipelines and ensuring backward-compatible API changes. Colleagues know him for cleaning up subtle numerical and shape bugs that improve robustness—work that often goes unnoticed but materially stabilizes ML systems.
11 years of coding experience
6 years of employment as a software developer
PhD Candidate (withdrawn), Computer Science, PhD Candidate (withdrawn), Computer Science at Nara Institute of Science and Technology
Github Skills (19)
lib10
python10
expressions10
abstract-syntax-tree10
machine-learning10
math-parser10
latex10
code-generation10
expression10
tensorflow10
parsing10
parse10
nlp10
computer-engineering9
refactoring8
Programming languages (11)
Github contributions (5)
google/latexify_py
Jul 2020 - Dec 2022
A library to generate LaTeX expression from Python code.
Role in this project:
Back-end Developer
Back-end DeveloperContributions:13 releases, 214 reviews, 98 commits in 2 years 5 months
Contributions summary:Yusuke appears to be the primary developer of the `latexify_py` library. They started the project, adding the core `latexify.py` file with the initial implementation to translate Python code into LaTeX. They subsequently fixed bugs related to unary operations, added support for additional mathematical functions, and introduced set operations. Later, they refactored the code to use `expression_rules`. The user also made several enhancements to improve the usability of the library by fixing a bug related to removing the multiply operator and adding support to remove the function name for the expression code.
python-codepythonlatex
tensorflow/lingvo
Jul 2019 - Jul 2020
Lingvo
Role in this project:ML Engineer
ML EngineerContributions:11 commits in 11 months
Contributions summary:Yusuke primarily focused on refining the GMM Attention implementation within the Lingvo framework. Their contributions involved modifying the shapes of intermediate values, ensuring consistency in reshaping operations, removing redundant code (e.g., `tf.to_float()`), adding shape annotations, and introducing an option for backward compatibility. These changes suggest a deep understanding of the attention mechanism and the underlying TensorFlow framework. The user also contributed to defining and integrating TFDataInput, further enhancing the data pipeline capabilities of the project.
asrtranslationctcspeech-recognitiontensorflow
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.